For the complete documentation index, see llms.txt. This page is also available as Markdown.

projects_import

Create a new draft from a parsed exported workflow_json object, optionally overriding project_name and folder_id.

Effect: This call saves data or changes configuration. Review the behavior below before using it.

Inputs

Name
Type
Requirement
Description

workflow_json

object

Required

Complete parsed StackAI workflow export, as an object. Do not pass a JSON string or a signed file URL.

project_name

string

Optional

Optional name for the imported project; otherwise use the exported name.

folder_id

string

Optional

Optional destination folder ID. If omitted, the default destination is used.

Example

These are MCP tool arguments, not an HTTP request body. Replace bracketed IDs with values returned from your account.

Build the arguments from your complete exported file; this preserves fields that a hand-written minimal export could omit.

import json
from pathlib import Path

arguments = {
    "workflow_json": json.loads(Path("workflow.json").read_text()),
    "project_name": "Imported workflow",
}

Pass arguments to projects_import in your MCP client.

Result

Abbreviated result example; unshown fields are omitted for clarity.

Behavior

This does not edit an existing project and is not idempotent.

Keep the serialized export below roughly 200 KiB and within the transport envelope.

Validate the draft and resolve connections for the target organization before publishing.

Troubleshooting

For an oversized export, use project import in the StackAI interface. File upload tools store bytes and do not import a workflow. After an uncertain response, inspect projects before importing again.

Projects and workflow authoring

Connection, permissions, pagination, and recovery

Was this helpful?